The “AI×UX Explorer” – From AI Theatre to UX Magic #UXCE25

How to systematically develop AI features that solve real problems – instead of just signaling innovation – like website chatbots, that too often fail to understand the user's request.

Most AI implementations fail not technically, but experientially. This presentation introduces the “AI×UX Explorer”, a practical toolkit that helps teams navigate the jungle of AI possibilities strategically and identify successful AI use cases to avoid costly tech experiments.

Experience a fresh perspective to AI – completely non-technical and tangible: Discover the menagerie of AI creatures and their capabilities as well as how and where we can make use of them to improve user/customer experiences (and therefore propel the business).

The framework provides immediately applicable tools across multiple relevant and often overlooked dimensions – For teams ready to move beyond AI theatre, this framework provides the systematic approach to identify, design and implement AI that truly enhances user experience.

#AI #ArtificialIntelligence #UX #UserExperience #CX #CustomerExperience #AIDesign #ProductStrategy #ServiceDesign #DigitalTransformation #AIIntegration #HumanCenteredAI #Innovation #Framework #Toolkit #Canvas #Carddeck #UXCE25 #UXCampEurope

Transcript

1.   THE AI×UX EXPLORER:    FROM AI THEATRE TO UX MAGIC  UX CAMP EUROPE, BERLIN JUNE 8, 2025 BENNO LŒWENBERG
2. Sorry, I didn’t understand that. Could you please rephrase your request?
3.   THE SHOWCASE PROBLEM  —Sarah’s emails stop syncing with iPhone (22:30, urgent presentation tomorrow) —the only support option available is that chat bot (simplified example)
4.   THE AI×UX CHALLENGE  “Most AI doesn’t fail technically. It fails experientially.”
5.   THE UX×AI CHALLENGE  “Building fast is easy now [with AI]. Building what people want is still hard!” Source: Caitlin Sullivan
6.   WHAT IF …  + AI seamlessly enhanced the customer/user experience? + we systematically identified the right AI use cases for great UX?
7.   THE AI×UX EXPLORER 
8.   CLARITY IN THE AI JUNGLE  A systematic AI×UX expedition: Pinpoint exactly where and how AI can enhance UX challenges or unlock new possibilities.
9.   WHERE DOES THE UX CHALLENGE LIE? 
10.   CONTEXT LEVELS 
11.  Pond    Interaction Model (UI element)   River    User Flow (User task)   Plateau    User Journey (User goal)  Graphic: Benno Lœwenberg   CONTEXT LEVELS 
12. Find UX challenges at each level, solve them at the right level   CONTEXT LEVELS FOR AI×UX  Graphic: Benno Lœwenberg
13. Graphic: Benno Lœwenberg Key Questions: How can AI enhance this specific interaction? How visible should the AI be at this point? What feedback does the user need here?   LEVEL 1: INTERACTION POINT 
14. Graphic: Benno Lœwenberg Key Questions: Where in the sequence does AI add most value? How can AI reduce steps in the flow? Can AI anticipate the next step?   LEVEL 2: USER FLOW 
15. Graphic: Benno Lœwenberg Key Questions: Which parts of the journey create friction? How can AI connect previously siloed parts? What data across the journey can improve AI performance?   LEVEL 3: USER JOURNEY 
16. —Problem Location: Mobile app (limited screen, hard to diagnose) —Solution Location: Desktop browser (optimal context) —Challenge: Traditional chatbots fail – no cross-device intelligence   SHOWCASE: LOCATION  (simplified example)
17.   WHAT KIND OF AI    DO WE WANT TO USE? 
18. 1 1   SHOW BREED 
19.   SHOW BREED (AI THEATRE)  —Attention-grabbing but disconnected —Chatbot icons sitting idle in corners —“AI-powered” badges and marketing —Features that require new behaviors —High initial curiosity, low sustained use
20.   ECOSYSTEM SPECIES 
21.   ECOSYSTEM SPECIES (AI-NATIVE)  + Deeply integrated, solving real problems + Netflix: Personalized thumbnails + Spotify: Discovery playlists + Google Photos: Smart organization + Apple: Face ID authentication
22.   FROM SURFACE TO SUBSTANCE  Designing and creating exceptional AI experiences that truly matter — from surface features to fully integrated solutions.
23.   WHICH AI CREATURES    CAN SUPPORT UX? 
24.   DISCOVERING THE AI MENAGERIE  The AI jungle teems with more creatures than first appears — each with unique superpowers for enhancing UX. The art lies in knowing which creature to summon for which UX challenge.
25.   THE EMOTIONS READER 
26.   THE EMOTIONS READER  + Capabilities: Emotion & mood detection from text, voice & facial expressions + Behavior: Responds to user frustration, adapts tone accordingly + Use Cases: User Insights, Customer Satisfaction Monitoring Example: Detects frustrated chat messages → escalates to human support
27.   THE CONTEXT ADAPTER 
28.   THE CONTEXT ADAPTER  + Capabilities: Situation-based interface adaptation + Behavior: Learns from usage patterns, adapts proactively + Use Cases: Individual UX, Adaptive Interfaces Example: Work hours → Business mode, Evenings → Relaxed mode
29.   THE PROCESS GUIDE 
30.   THE PROCESS GUIDE  + Capabilities: Multi-step workflow guidance with goal orientation + Behavior: Anticipates next steps, keeps users on track + Use Cases: User Flows + Customer Journeys Example: Onboarding process → guides through 5 steps, remembers progress
31.   THE ISSUE RESOLVER 
32.   THE ISSUE RESOLVER  + Capabilities: Natural language problem analysis & solution suggestions + Behavior: Cross-device intelligent, learns from solution success + Use Cases: Self-Service Support, Technical Troubleshooting Example: eMail sync problem → analyzes, suggests solutions
33. Graphic: Benno Lœwenberg   AI SPECIES   LINEAGE  Specialized (Domain-specific) FinTech 4 HealthTech 4 IndustrialTech 4 Language Models Machine Learning Hybrid (Multi Modal) Computer Vision Interactive 5 (Conversational) Predictive 5 Contextual 4 (Context-Aware) Emotion Recognition 3 Creative 5 (Generative) Perceptive 5 (Pattern Recog.) Emotion Recognition 1 Recognition 5 Knowledge Processing 4 Personalization 4 Intelligent Automation 4 Guidance 4 Creative-Visual 3 (Gen. Vision) Issue Resolver Protecting 4 (Security & Trust) Audio Sound 4 (Audio Processg.) Communicating 2 (Voice Interface) Autonomous Systems 3 Agentic AI 3 Context Adapter Emotion Reader Process Guide
34. Problem vs. Capability Matching + Map user problems to AI creature capabilities + Assess fit quality: High/Medium/Low + Select optimal matching option(s)   ALIGN AI CAPABILITY WITH USER NEEDS 
35. + Why perfectly suited: LLM Foundation: Understands natural language problem descriptions + Cross-Context: Recognizes problems across devices + Solution-Oriented: Focus on practical, actionable next steps + Adaptive Communication: Adjusts to user expertise level → The Issue Resolver combines problem understanding + solution competence + cross-device intelligence   SHOWCASE: ISSUE RESOLVER  (simplified example)
36.   HOW SHOULD THE HUMAN-AI    COLLABORATION LOOK LIKE? 
37.   AI SUPPORTS 
38.   AI COLLABORATES 
39.   AI ACTS AUTONOMOUSLY 
40.   AI AUTONOMY LEVELS  1. AI informs — AI suggests, human decides 2. AI supports — AI actively assists the human 3. AI collaborates — AI and human work together 4. AI executes — AI does it, human supervises 5. AI decides — AI operates completely autonomously
41.   AI AUTONOMY LEVELS  Graphic: Benno Lœwenberg 100 % Human AI 0 % 1. informs 2. supports 3. collaborates 4. executes 5. decides Human AI Shared responsibility AI dominates
42.  EXAMPLES  1. Dashboards, reports, analytics 2. Intelligent autocorrect, appointment suggestions 3. Co-design, collaborative problem-solving 4. Automated customer service, routine processes 5. Algorithmic trading, predictive maintenance
43.   UX-FACTOR: OVERRIDE  Graphic: Benno Lœwenberg informs supports collaborates executes decides 100 % Human AI 0 % Human AI The ability to stop or override the AI at any time
44.   UX-FACTOR: HUMAN IN THE LOOP  Graphic: Benno Lœwenberg 100 % Human AI 0 % Human AI The human actively works together with the AI Little collaboration Human excluded informs supports collaborates executes decides
45. + AI handles: Problem detection, analysis, solution suggestions + Human decides: Which solution to try, when to escalate + User feels empowered, not replaced + AI accelerates problem-solving + Fallback to human support always available   SHOWCASE: UX SWEET-SPOT   (simplified example)
46.   HOW PROMINENT SHOULD THE AI BE? 
47.  HI-VIZ 
48.   DELIBERATELY VISIBLE AI  + When transparency builds trust + For complex decisions + For active user control
49.   HIDDEN MAGIC 
50.   SEAMLESSLY INTEGRATED AI  + When efficiency takes priority + For routine, low-risk tasks + For smooth workflows
51.   SHOWCASE: UI PART  × × (simplified example)
52. When users have problems, they want to understand how they’re being helped + User understands the AI’s reasoning + Multiple solution paths are transparent + User can choose alternative approaches + Builds confidence in AI recommendations   SHOWCASE: UI PART 
53. + Solve real user pain points — the best experiences feel like pure magic + Technology vanishes, user delight remains + Invisible integration beats flashy features + Enhance existing workflows, don’t create new ones   THE AI×UX MAGIC FORMULA 
54.   THE AI×UX EXPLORER 
55.   THE AI×UX EXPLORER TOOLS 
56. Module Designer Shape Customer and User Experience Module Builder Enable Scaling and Growth Module Explorer Identify Usecases for AI Graphic: Benno Lœwenberg   THE AI×UX SYSTEM 
57. Want to use this framework to solve your UX challenge too?
58.   BENNOLŒWENBERG  New Momentum for AI, Innovation and Customer Experiences. Simple and Effective. website linkedin